On 15 August 2026, the 80th Independence Day, the Prime Minister devoted a section of his Red Fort address to electricity. Chips, artificial intelligence and data centres, he said, cannot run without enormous amounts of power. Energy security is the need of the hour. One means of ensuring it is nuclear energy: 100 gigawatts of nuclear capacity by 2047, five new reactors operational within this decade, and, he added, a milestone achieved this year in breeder technology that takes India towards self-reliance in nuclear fuel. Three days earlier, in reply to Starred Question No. 360 in the Lok Sabha, the Department of Atomic Energy had set out where the programme stands. India has 24 reactors, with an installed capacity of 8,780 MW.
The share of nuclear energy in electricity generation in 2025-26 was 3.1%. Set the speech beside the answer. Seven decades of the atomic energy programme have yielded 8.78 GW. The next 21 years are meant to yield another 91 GW or so, roughly ten times the existing stock. Whether that happens will depend less on physics than on statutes, rules, sanctions and mines. Begin with what is being done, because a great deal is. The Nuclear Energy Mission, announced in the Union Budget 2025- 26, carries a provision of Rs 20,000 crore for small modular reactors (SMRs), with at least five indigenous SMRs to be operational by 2033. The Bhabha Atomic Research Centre is developing three designs: a 220 MWe Bharat Small Modular Reactor and a 55 MWe SMR, both sited at Tarapur, and a high temperature gas cooled reactor of up to 5 MWth at Visakhapatnam, meant to be coupled with a thermochemical cycle for hydrogen. The special reactor vessel alloy, ApuRVA, has been developed with Indian industry.
On 6 April 2026, the Prototype Fast Breeder Reactor at Kalpakkam attained criticality, the first physical demonstration of the second stage of Homi Bhabha’s three-stage programme, the stage that eventually unlocks thorium. The roadmap laid before the Lok Sabha on 22 July 2026 (Unstarred Question No. 528) is quite explicit. About 22 GW by 2031-32 from projects under construction and sanction, about 54 GW by 2047 from NPCIL, and the balance 46 GW from other public enterprises, state governments, joint ventures and the private sector. That last clause is where the reform lies. The Atomic Energy Act, 1962 reserved nuclear power generation for the state. The Sustainable Harnessing and Advancement of Nuclear Energy for Transforming India (SHANTI) Act, 2025 replaces that architecture with a single statute.
Licensed private participation in non-strategic activities, a statutory Atomic Energy Regulatory Board with its own regulation-making and enforcement powers, operator-led liability with graded caps, supplier liability governed by contract, and mandatory financial provisioning for decommissioning at the approval stage. Enrichment beyond a notified threshold, reprocessing and the accounting of nuclear material remain with the state. Draft rules were circulated for public consultation on 14 August 2026, providing for approval in principle before licence, stage-wise regulatory scrutiny, and admission of foreign reactor designs with proven operating records. The colonial default of prohibited unless permitted has been replaced by permitted unless prohibited, with a defined negative list.
That is the correct architecture. Legislation, however, is the easy part. Consider what stands between announcement and outcome. First, the sanctioning drought. The DAE told the Lok Sabha on 29 July 2026 (Unstarred Question No. 1775) that no new nuclear power project was sanctioned in the last three years; the achievements listed, KAPP-3 and 4 and RAPP-7, adding 2,100 MW, were completions of old sanctions. The same Department states that an SMR needs 60 to 72 months from administrative and financial sanction to establishment. Five by 2033 is therefore not a 2033 decision. It is a 2026 one. Second, the arithmetic of pace and money. Reaching 100 GW from 22 GW in 2032 implies average additions of about 4.5 GW every year for fifteen years, against 2,100 MW added in the last three.
The Energy and Resources Institute, in India’s Nuclear Energy Vision: Strategic Pathways for SMR Deployment (2026), estimates the investment requirement at Rs 23 to 25 lakh crore, about Rs 1.0 to 1.2 lakh crore a year, and notes that a base capital cost of Rs 12 to 14 crore per MW becomes Rs 20 to 25 crore once interest during construction, escalation and contingency are counted. Time overruns are not a scheduling nuisance; they are the single largest cost. Third, fuel. Domestic uranium production is about 600 tonnes a year; 18,842.60 MT of uranium has been imported since 2008-09 for safeguarded reactors. A parliamentary committee on regulatory enablement for strategic minerals has pointed out that the Mines and Minerals (Development and Regulation) Act does not permit contiguous extensions of a mining lease when an ore body runs beyond its boundary, forcing fresh leases and fresh clearances, a real problem for deposits below 600 metres, as at Narwapahar.
It recommended singlewindow clearances with statutory timelines and a DAE-chaired coordination platform. The government’s final response is awaited. There is a fourth constraint, rarely mentioned. People. The key public sector units employ about 36,076 persons. TERI’s estimate of peak construction manpower for the build-out is 120,000 to 200,000, with 38,000 to 60,000 needed in operations, besides welders, inspectors and regulators the present training pipeline was never sized for. And the reform itself is unfinished. The DAE’s own answers state that the SHANTI Act’s provisions take effect only upon notification, and that states, industries and private firms can come in once rules are framed.
TERI’s report lists what the Act leaves to subordinate instruments: standard ownership and operating models, siting and land-use frameworks adapted to small reactors, financing structures for first-of-a-kind projects, fuel supply and backend management, and the scaling of regulatory capacity. Each is a rule-making task, not a legislative one, which is precisely why each is easy to defer. None of this is an argument against the target. India runs one of the cheapest nuclear programmes in the world, at 47.64 dollars per MWh, lower than the United States, France and China. Competence is not the constraint. Throughput is. The throughput of sanctions, clearances, leases, rules and trained hands. In a reactor, criticality is the point at which the chain reaction sustains itself, without external prodding. The PFBR reached it on 6 April 2026. The programme has not, yet.
